Investigation of machine parameters on the surface quality in routing soft maple. (Solid Wood Products).
Furniture manufacturers are currently using relatively high quantities of soft maple. Machining defects such as torn or fuzzy grain often occur in soft maple parts at the computer numerical control (CNC) router, resulting in machine-room yield loss. This paper explores the machining quality of soft maple on end grain, flat side grain, and curved side grain. Using a CNC router, machined surface quality was evaluated as a function of feed speed, spindle speed, tool direction, and tool wear. On flat side grain surfaces, climb cutting produced a higher quality surface compared to conventional cutting. Conventional cutting, on the other hand, produced a better quality surface than climb cutting on end grain. The effect of tool wear was not consistent; however, differences were observed when the other cutting parameters were optimum for producing a high quality surface. End grain and curved side grain surfaces exhibited a decreased machining quality with increasing feed per knife. Flat grained surfaces, which gener ally machine better than cross grain or end grain, showed no difference in quality when climb cut, and little change with conventional cutting.
Soft maple is often employed as a less expensive substitute for hard maple species in furniture and cabinet manufacturing. The machining quality of soft maple, however, is sometimes poor due to cross grain and the presence of tension wood (7,8). Wood parts that develop torn or fuzzy grain while being machined by computer numerical control (CNC) routers are often rejected because the visibility of the defect will increase with subsequent finishing processes, and cannot easily be remedied by sanding. Improvements in the machining techniques of soft maple would not only improve machine-room yield, it would also encourage the expanded utilization of the species.
Work by McKenzie (4,5) has shown that although other published works claim an improved surface quality with increasing cutting speed or rpm, it is usually the accompanying change in feed per knife, or chip progression, that results in an improved surface quality. Feed per knife is the inverse of knife marks per inch, and is defined as:
Feed per knife = feed rate X 12/rpm X number of knives
where feed per knife is expressed in inches and feed rate is in feet per minute. McKenzie has shown that when the feed per knife remains the same, very little difference in surface quality occurs. Davis (2) similarly suggested that machining quality in planing is governed largely by knife marks per inch (inverse of feed per knife) and not by cutterhead rpm alone. Although both cutting speed and feed per knife are varied, the changes in the feed per knife will be the parameter that controls the surface quality of the specimen.
The objective of this work was to define the machining quality of soft maple (soft maple includes both red and silver maple) as a function of cutting parameters and tool dullness, specifically as it relates to routing.
MATERIAL AND METHODS
Although ASTM D 1666 (1) served as a guide, the procedures were modified, as suggested by the standards, when it was decided by the principal investigators that improvements or adjustments were necessary. For example, ASTM D 1666 was written specifically to evaluate different species requiring 50 test samples for each species tested. A much smaller group sample size was studied for the current work. However, sample selection and specimen conditioning and preparation were conducted according to ASTM standards, with some exceptions, which will be noted.
Approximately 350 board feet of kilndried 4/4 12-feet-long No.1 Common soft maple was obtained for the machining study. The lumber was planed and sawn to produce samples that were 0.75 inch thick, 3 inches wide, and 12 inches long. The lumber was not clear, nor were the samples cut from the lumber totally free from growth characteristics that would be considered defects. Small knots, mild cross grain, and stain were allowed if it was judged that the sample could be machined without encountering these areas. Before routing, the sample blocks were conditioned to a constant weight in a 6 percent equilibrium moisture content chamber.
Samples were machined to the pattern shown in Figure 1 using two passes. The first pass was done with a two fluted downspiral hogging bit 0.5 inch in diameter. The finishing cut was 1/16 inch deep, and was done with a straight profile carbide knife using a 0.75-inch diameter single flute insert tool bit. The tungsten carbide blade had a 3 percent cobalt binder. The pattern employed is the same as in ASTM D 1666 for shaping, and allows the machining and evaluation of the end grain and side grain surfaces.
These machining tests were conducted with a CNC router (Thermwood Model 40 with a 5 ft. by 5 ft. bed) with a carbide-tipped tool rather than the several machining methods described by ASTM D 1666.
The test design considered different degrees of tool wear and tool speed. The cutting speed was accomplished by varying the feed speed and the spindle speed on the CNC router to obtain a range of feed per knife. Table 1 shows the five feed speeds in inches per minute (ipm) and two spindle speeds in revolutions per minute (rpm) used to obtain the feed per knife or chip progression shown. These cutting conditions provided a broad range of chip progression over which to evaluate the surfaces produced.
For each of the feed speed combinations shown in Table 1, two levels of tool wear were evaluated. Degree of tool wear has been previously described by Sheikh-Ahmad et al. (6). Using their classification, the two levels of tool wear employed throughout all feed speed combinations were sharp (nose width = 5 [micro]m) and slightly worn (nose width = 40 [micro]m). An additional level of tool wear, moderately worn (nose width = 80 [micro]m), was employed for two feed speed combinations: 200 and 300 ipm at 18,000 rpm.
For each of the eight feed speed combinations evaluated in this study, the different wood surfaces machined and the tool directions were considered. The three separate surfaces evaluated were end grain, flat side grain, and curved side grain, as indicated in Figure 1. The different machined grain surfaces were evaluated separately. Since both climb cutting and conventional cutting (tool directions) can occur in a router operation, it was an important effect to evaluate. Climb and conventional cutting are illustrated in Figure 2.
SURFACE QUALITY EVALUATION
Three individuals visually graded each piece of machined soft maple using the quality standards described by D 1666: Grade 1 = excellent; Grade 2 = good; Grade 3 = fair; Grade 4 = poor; Grade 5 very poor.
Defect type and severity were also classified according to raised, fuzzy, and torn grain. The descriptive grade then assigned to the surface of each piece was the largest average (of three graders) of the three defect types. This descriptive grade can be considered the limiting quality characteristic for that sample surface since it represents the worst quality for that particular surface (whether due to raised, fuzzy, or torn grain). In a practical sense, grades 1 and 2 represent surfaces that would be acceptable in a manufacturing environment, while grades 3 and higher would require excessive sanding and rework.
RESULTS AND DISCUSSION
The mean descriptive sample grades determined for each of the main effects are listed in Table 2. Main effect means were separated using the Student-Newman-Keuls procedure (0.05 level). Means with the same letter are not significantly different one from the other.
The mean descriptive quality grades for each grain surface were: end grain = 2.45; flat side grain = 1.43; and curved side grain = 1.79. Each was significantly different from the other two surfaces. End grain exhibited the worst surface quality, while the flat side surface grain represented the best quality. Because of the expectation that different wood surfaces will behave very differently during machining and that this will be reflected in the resulting surface quality, each grain surface was considered separately in the analysis.
Significant differences in feed speed were found in machining end grain and curve grain surfaces over the range from 100 to 600 ipm. No differences in feed speeds for flat grain surfaces existed, and on average, the quality of the flat grain surfaces generated was acceptable (grade 2 or better). For end grain, feed speeds of 300 ipm and higher generated unacceptable surfaces (higher than grade 2). For curved grain surfaces, unacceptable surfaces were generated only at 600 ipm.
Two spindle rotational speeds were studied. No significant difference in mean surface grade attributable to spindle speed (rpm) alone was found for any of the grain surfaces, as indicated in Table 2.
The direction of feed can have a significant impact on the quality of the machined surface. For solid wood, climb cutting (also called downmilling) is generally recommended over conventional cutting (also called upmilling) (Fig. 2). In this study, tool direction was a significant factor when machining the end and flat grain surfaces. Climb cutting produced a significantly higher quality surface compared to conventional cutting on the flat grain surface. Surprisingly, this trend was reversed for end grain surfaces, with conventional cutting producing a significantly better quality surface. No difference between climb and conventional cutting on the curved grain surface existed.
When considered as a main factor, differences due to tool wear were present but were not consistent (Table 2). Further analysis found significant differences due to tool wear for the conventionally machined end grain and climb cut flat grain surfaces (Table 3) for samples machined at 18,000 rpm. At most feed speeds for these two instances, tools that were slightly or moderately worn produced surfaces of poorer quality than did sharp tools. As mentioned in the previous section, conventionally machined end grain and climb cut flat grain surfaces produced significantly higher quality grades than those produced with the opposite tool direction. Possibly, tool wear differences could only be detected when other quality inhibiting parameters were minimized. Curved grain surfaces did not show the same degree of differences due to tool wear as did the other two surfaces.
CHIP PROGRESSION OR FEED PER KNIFE
Surface quality differences that existed due to change in chip progression, or feed per knife, were examined. The numerical quality grade data are presented in Table 4, separated by grain, direction, and tool wear. Significant differences in mean quality grades between groups having different feeds per knife are also indicated in Table 4. The relationship between feed per knife and surface quality is also summarized in Figure 3.
Increasing feed per knife for end grain surface machining resulted in a significant loss in quality regardless of cutting direction or tool wear, as measured by the numerical increase in the quality grade. When climb cutting end grain, acceptable quality surfaces (quality grade of 2 or less) were produced, on average, with a feed per knife of 0.0111 inch or less. An acceptable quality surface could be produced when conventional cutting up to a feed per knife of 0.0167 inch. In addition, climb cutting resulted in a much larger loss in quality with increasing feed per knife than did conventional cutting (although both cutting methods saw significant quality loss). This is counter to the conventional wisdom that climb cutting should be used when machining wood. The significant decrease in quality with increasing feed per knife for climb and conventional cut end grain is illustrated in Figure 3.
In climb cutting flat grain surfaces, increasing the feed per knife had no significant impact on quality (Table 4 and Fig. 3). No difference in surface quality attributable to changing feed per knife when flat grain surfaces were climb cut (with either sharp or dull knives) is illustrative of the ease with which flat grain surfaces can be machined.
Increasing the feed per knife for curved grain surfaces resulted in a significantly lower surface quality, regardless of cutting direction or tool wear. Whether climb or conventional cutting, an acceptable quality surface was produced, on average, with a feed per knife of 0.0167 inch or less (Table 4 and Fig. 3).
SUMMARY AND CONCLUSIONS
Not surprisingly, the best machined quality was exhibited by the flat side surface grain while the worst surface quality occurred on the end grain. Machining quality on curved grain surfaces was intermediate.
Tool direction was a significant factor in determining surface quality for end grain and for flat side grain, but not for the curved side grain surfaces. On flat grain surfaces, climb cutting produced a higher quality surface compared to conventional cutting. This was reversed for end grain surfaces, with conventional cutting producing a better surface than climb cutting.
Although the effect of tool wear was not consistent, differences were observed when other cutting parameters were optimum for producing a high quality surface. Specifically, tools with slight or moderate wear produced lower quality surfaces on flat side grain during climb cutting. Similarly, a lower surface quality resulted when slight or moderately worn tools were used in the conventional cutting of end grain surfaces.
Increasing the feed per knife, or chip progression, resulted in a decreased machining quality for end grain and curved side grain surfaces. The maximum feed per knife that, on average, generated acceptable quality end grain surfaces (quality grade of 2 or less) was 0.0111 inch or less for climb cutting, and 0.0167 inch or less for conventional cutting. For curved grain surfaces, a feed per knife of 0.0167 inch or less resulted in acceptable quality for both climb and conventional cutting. Quality machining of flat grain surfaces can be accomplished with ease, as evidenced by the lack of differences in surface quality with increasing feed per knife on flat grain surfaces with either sharp or dull knives.
In conclusion, what this study means is that in order to obtain satisfactory surfaces when machining soft maple, the operator must take several parameters into consideration. The first is obviously to insure sharp blades when machining soft maple. The second is to use conventional cutting when machining across the grain and climb cutting when machining in other grain directions. In addition, this study confirms Mackenzie's finding that surface quality is controlled by the calculated feed per knife. Specifically, to insure quality machined end grain surfaces, the operator should limit the feed per knife to 0.0111 inch per tooth or less for climb cutting, and 0.0167 inch per tooth for conventional cutting when machining end grain. When machining other grain directions the feed per knife should be limited to 0.0167 inch per tooth for both climb and conventional cutting. These recommendations for feed per knife are in the range recommended by Effner (3) to produce quality surfaces and can be used to calculate fee d rates for CNC operations that have different machining parameters than those studied.
[FIGURE 3 OMITTED]
TABLE 1 Combination of feed speeds and spindle speeds used in this study, and the corresponding feed per knife. Spindle speed 12,000 rpm 18,000 rpm Feed speed Feed per knife (ipm) (in.) 100 0.0083 0.0056 200 0.0167 0.0111 (a) 300 0.0250 0.0167 (a) 450 -- 0.0250 600 -- 0.0333 (a)An additional level of tool wear was employed for these two feed combinations. TABLE 2 Comparison of mean surface quality grades for each of the main test effects. (a) END GRAIN Feed (ipm) 100 200 300 450 600 Mean 1.10 A 1.84 B 3.14 C 3.62 D 3.74 D n 64 96 96 32 32 RPM 12,000 18,000 Mean 2.51 A 2.43 A n 96 224 Direction Conventional Climb Mean 1.96 A 2.94 B n 160 160 Tool wear Sharp Slight Moderate Mean 2.29 A 2.66 B 2.43 A n 160 128 32 FLAT GRAIN Feed (ipm) 100 200 300 450 600 Mean 1.35 A 1.30 A 1.54 A 1.59 A 1.53 A n 64 95 91 32 27 RPM 12,000 18,000 Mean 1.34 A 1.47 A n 93 216 Direction Conventional Climb Mean 1.58 A 1.29 B n 149 160 Tool wear Sharp Slight Moderate Mean 1.34 A 1.49 AB 1.69 B n 156 124 29 CURVED GRAIN Feed (ipm) 100 200 300 450 600 Mean 1.24 A 1.48 A 2.04 B 2.00 B 2.84 C n 64 96 96 32 32 RPM 12,000 18,000 Mean 1.74 A 1.81 A n 96 224 Direction Conventional Climb Mean 1.87 A 1.70 A n 160 160 Tool wear Sharp Slight Moderate Mean 1.70 A 1.87 A 1.91 A n 160 128 32 (a) Means followed by the same capital letter within the same row are notsignificantly different at the 0.05 level using Student- Newman-Keuls procedure. TABLE 3 Mean surface quality grades for samples machined at 18,000 rpm by degree of tool wear for each surface grain, cutting direction, and feed speed. (a) Feed Slight Moderate speed Sharp wear wear END GRAIN SURFACE Climb cutting 100 1.04 A 1.00 A 200 1.25 A 1.46 A 1.50 A 300 3.44 A 3.21 A 3.08 A 450 4.92 A 4.96 A 600 4.83 A 4.96 A Conventional cutting 100 1.00 A 1.17 A 200 1.60 A 1.71 A 1.92 A 300 1.71 A 2.21 A 3.21 B 450 1.75 A 2.87 B 600 2.04 A 3.16 B FLAT SIDE GRAIN SURFACE Climb cutting 100 1.17 A 1.50 A 200 1.00 A 1.04 A 1.79 B 300 1.10 A 1.71 AB 2.33 B 450 1.00 A 1.67 B 600 1.00 A 1.46 B Conventional cutting 100 1.67 A 1.38 A 200 1.46 A 1.33 A 1.33 A 300 1.73 B 1.71 B 1.10 A 450 1.66 A 2.04 A 600 2.22 A 1.67 A CURVED SIDE GRAIN SURFACE Climb cutting 100 1.04 A 1.21 A 200 1.06 A 1.25 A 1.75 B 300 1.52 A 1.88 A 2.33 A 450 1.83 A 1.96 A 600 3.00 A 2.75 A Conventional cutting 100 1.29 A 1.54 A 200 1.54 A 1.38 A 1.79 A 300 1.81 A 2.13 A 1.75 A 450 2.21 A 2.00 A 600 2.79 A 2.83 A (a)Means followed by the same capital letter within the same row are not significantly different at the 0.05 level using the Student-Newman-Keuls procedure. TABLE 4 Mean surface quality grades by size of feed per knife (chip progression) for each grain surface, tool direction, and tool wear. (a) Feed per knife (in.) Tool 0.0056 0.0083 0.0111 wear END GRAIN SURFACE Climb cutting Sharp Quality 1.04 A 1.00 A 1.25 A n 8 8 16 Slight Quality 1.00 A 1.33 AB 1.46 B n 8 8 8 Conventional Cutting Sharp Quality 1.00 A 1.00 A 1.60 AB n 8 8 16 Slight Quality 1.17 A 1.29 A 1.71 AB n 8 8 8 FLAT SIDE GRAIN SURFACE Climb cutting Sharp Quality 1.17 A 1.00 A 1.00 A n 8 8 16 Slight Quality 1.50 A 1.25 A 1.04 A n 8 8 8 Conventional cutting Sharp Quality 1.67 B 1.08 A 1.46 AB n 8 8 16 Slight Quality 1.38 A 1.79 A 1.33 A n 8 8 8 CURVED SIDE GRAIN SURFACE Climb cutting Sharp Quality 1.04 A 1.00 A 1.06 A n 8 8 16 Slight Quality 1.21 A 1.12 A 1.25 A n 8 8 8 Conventional cutting Sharp Quality 1.29 AB 1.08 A 1.54 AB n 8 8 16 Slight Quality 1.54 A 1.63 A 1.38 A n 8 8 8 Feed per knife (in.) Tool 0.0250 0.0167 0.0330 wear END GRAIN SURFACE Climb cutting Sharp 3.29 B 4.94 C 4.83 C 24 16 8 Slight 3.21 C 4.94 D 4.96 D 16 16 8 Conventional Cutting Sharp 1.65 AB 2.23 B 2.04 B 24 16 8 Slight 2.15 B 2.96 C 3.13 C 16 16 8 FLAT SIDE GRAIN SURFACE Climb cutting Sharp 1.07 A 1.15 A 1.00 A 24 16 8 Slight 1.48 A 1.44 A 1.46 A 16 16 8 Conventional cutting Sharp 1.65 B 1.76 B 2.22 C 24 14 6 Slight 1.56 A 1.82 A 1.67 A 16 15 5 CURVED SIDE GRAIN SURFACE Climb cutting Sharp 1.42 A 2.17 B 3.00 C 24 16 8 Slight 1.67 AB 2.27 ABC 2.75 C 16 16 8 Conventional cutting Sharp 1.82 ABC 2.27 BC 2.79 C 24 14 6 Slight 1.98 AB 2.17 AB 2.83 B 16 16 8 (a)Means followed by the same letter within the same row are not significantly different at the 0.05 level using the Student- Newman-Keuls procedure.
(1.) American Society for Testing Materials. 1996. Standard methods for conducting machining tests of wood and wood-base materials. D-1666-87 (Reapproved 1994). ASTM, West Conshohocken, PA.
(2.) Davis, E.M. 1962. Machining and related characteristics of United States hardwoods. Tech. Bull. No. 1267. USDA, Washington, DC. 65 pp.
(3.) Effner, J. 1992. Chisels on a Wheel: A Comprehensive Reference to Modern Wood-working Tools and Materials. Prakken Pub., Inc. Ann Arbor, MI. 200 pp.
(4.) McKenzie, W.M. 1960. Fundamental aspects of the wood cutting process. Forest Prod. J. l0(9):447-456.
(5.) _____ 1970. Feed per cut, a very important parameter in working wood-based materials. The Australian Timber J. pp. 133-135.
(6.) Sheik-Ahmad, J., J.S. Stewart, and J.A. Bailey. 1995. Performance of different PVD coated tungsten carbide tools in the continuous machining of particleboard. In: Proc. of the 12th Inter. Wood Machining Seminar. October, Kyoto, Japan. pp. 282-291.
(7.) USDA Forest Products Laboratory. 1965. Raised, loosened, torn, chipped, and fuzzy grain in lumber. Res. Note FPL-099. USDA Forest Prod. Lab., Madison, WI. 15 pp.
(8.) Wengert, E.M. and F.M. Lamb. 1994. A Handbook for Improving Quality and Efficiency in Rough Mill Operations. R.C. Byrd Hardwood Technology Center, Princeton, WV. 107 pp.
PHILIP H. MITCHELL *
RICHARD L. LEMASTER *
* Forest Products Society Member
The authors are, respectively, Assistant Professor and Wood Products Extension Specialist, North Carolina State Univ. (NCSU), Dept. of Wood and Paper Sci., Campus Box 8003, Raleigh, NC 27695-8003; and Research Director, NCSU, Wood Machining & Tooling Res. Program, Campus Box 8005, Raleigh, NC 27695-8005. This paper was received for publication in March 2001. Reprint No. 9291.
[C] Forest Products Society 2002.
Forest Prod. J. 52(6):85-90.
|Printer friendly Cite/link Email Feedback|
|Author:||Mitchell, Philip H.; LeMaster, Richard L.|
|Publication:||Forest Products Journal|
|Date:||Jun 1, 2002|
|Previous Article:||Wood-surface feature classification via extended-color imagery. (Fundamental Disciplines).|
|Next Article:||Comparison of strength properties and failure characteristics between fire-retardant-treated and untreated roof framing lumber after long-term...|